Vol. 47 No. 4/2000
1045–1060
QUARTERLY
Further studies on the role of phospholipids in determining
the characteristics of mitochondrial binding sites for type I
hexokinase.
Jan Hutny½ and John E. Wilson
Department of Biochemistry, Michigan State University, East Lansing, MI 48824-1319, U.S.A.
Received: 17 May, 2000; revised: 10 August, 2000; accepted: 29 August, 2000
Key words: hexokinase, mitochondria, mitochondrial binding, phospholipids
Previous work has indicated that two types (A and B) of binding sites for hexokinase
exist, but in different proportions, on brain mitochondria from various species.
Hexokinase is readily solubilized from Type A sites by glucose 6-phosphate (Glc-6-P),
while hexokinase bound to Type B sites remains bound even in the presence of
Glc-6-P. Type A:Type B ratios are approximately 90:10, 60:40, 40:60, and 20:80 for
brain mitochondria from rat, rabbit, bovine and human brain, respectively. The present study has indicated that MgCl2-dependent partitioning of mitochondrially bound
hexokinase into a hydrophobic (Triton X-114) phase is generally correlated with the
proportion of Type B sites. This partitioning behavior is sensitive to phospholipase C,
implying that the factor(s) responsible for conferring hydrophobic character is(are)
phospholipid(s). Substantial differences were also seen in the resistance of
hexokinase, bound to brain mitochondria from various species, to solubilization by
Triton X-100, Triton X-114, or digitonin. This resistance increased with proportion of
Type B sites. Enrichment of bovine brain mitochondria in acidic phospholipids
(phosphatidylserine or phosphatidylinositol), but not phosphatidylcholine or
phosphatidylethanolamine, substantially increased solubilization of the enzyme after
incubation at 37°C. Collectively, the results imply that the Type A and Type B sites are
located in membrane domains of different lipid composition, the Type A sites being in
domains enriched in acidic phospholipids which lead to greater susceptibility to
solubilisation by Glc-6-P.
The Type I isoenzyme of mammalian
hexokinase (ATP:D-hexose 6-phosphotrans.
ferase, EC 2.7.1.1) is ubiquitously expressed
in mammalian tissues [1] but found at particu-
This work was supported by NIH Grant NS 09910.
Address for correspondence: Jan Hutny, Department of Biochemistry, Agricultural University of
Wroc³aw, K.C. Norwida 31, 50-375 Wroc³aw, Poland; phone: (48 71) 320 5435; fax: (48 71) 321 1567,
e-mail: [email protected]
Abbreviations: VDAC, voltage dependent anion channels; Glc-6-P, glucose 6-phosphate; BSA, bovine serum albumin; SDS, sodium dodecyl sulfate.
½
1046
J. Hutny and J.E. Wilson
larly high levels in brain [2] where it plays an
important role in regulating the rate of cerebral glucose (Glc) metabolism [3]. The major
portion (> 80%) of the hexokinase activity in
brain is associated with mitochondria [4–9].
This physical proximity is thought to provide
the basis for functional interaction between
Glc phosphorylation by hexokinase and intramitochondrial ATP production by oxidative
phosphorylation, with resulting coordination
of the glycolytic and oxidative phases of Glc
metabolism [10, 11]. The outer mitochondrial
membrane protein responsible for specific interaction with hexokinase was originally isolated as the “hexokinase binding protein” [12]
but quickly shown [13, 14] to be identical with
porin (also called VDAC), the protein which
forms the transmembrane channel by which
metabolites enter or exit the mitochondria
[15]. It thus appears that hexokinase is directly associated with the physical structure,
the pore, by which substrate ATP may be delivered from its intramitochondrial source.
With mitochondria from rat brain, binding
of hexokinase is readily reversible, with about
90% of the enzyme being released by incubation of the mitochondria with mmol/l levels of
Glc-6-P [16], the product of the hexokinase reaction. However, this is not generally true for
mitochondrial hexokinases from brain of
other species, with a maximum of about: 60%,
40%, and 20% of the hexokinase being released by Glc-6-P with mitochondria from rabbit, bovine, and human brain [16], respectively. The hexokinase that is not released
with Glc-6-P can be solubilized by further
treatment of the mitochondria with high salt
concentrations [16, 17]. Previous work [16,
18] has led to the conclusion that this difference in sensitivity to release by Glc-6-P does
not result from intrinsic differences in the
hexokinase itself, but rather, reflects the existence of two different kinds of binding sites
for the enzyme. Thus, Type A sites are defined
as mitochondrial sites from which the enzyme
is released by the action of Glc-6-P, while Type
B sites are those at which hexokinase remains
2000
bound even in the presence of Glc-6-P. In these
terms, the Type A : Type B ratio of binding
sites is approximately 90 : 10, 60 : 40, 40 : 60,
and 20 : 80 for mitochondria from rat, rabbit,
bovine, and human brain, respectively.
Given the well-known heterogeneity of the
mitochondrial population in brain [19–22],
one obvious possibility was that the Type A
and Type B sites might reside on different
subpopulations of brain mitochondria. However, this is not the case, with both Type A and
Type B sites shown to coexist on the same mitochondria [18]. Isoforms of porin have also
been reported [14, 15, 23], and thus another
possibility was that Type A and Type B sites
might correspond to involvement of different
isoforms of porin in binding of hexokinase.
However, this also appears not to be the case
since rat and bovine brain mitochondria,
while differing markedly in ratio of Type
A : Type B sites, were indistinguishable in
their porin isoform content [18]. Thus, the
molecular basis for the occurrence of these
distinct types of hexokinase binding sites has
remained unclear. The present study suggests
that the phospholipid environment may play a
major role in determining the Type A-Type B
character of the mitochondrial binding site
for hexokinase.
MATERIALS AND METHODS
Materials. Bovine brain was obtained at the
Meats Laboratory (Michigan State University, U.S.A.), and transported to the laboratory on ice. Gross regions of white matter and
larger blood vessels were dissected away, and
the tissue was cut into smaller portions and
stored in liquid nitrogen. Brains from rats,
rabbits, and guinea pigs were obtained immediately after sacrifice of the animals and
stored in liquid nitrogen, as in previous studies from this laboratory [16, 18]. Triton X-114,
phospholipase C (catalog No. P-7633), and
phospholipids were obtained from Sigma
Chemical Co. (St. Louis, MO, U.S.A.), while
Vol. 47
Mitochondrial binding sites for hexokinase
Triton X-100 was from Research Products International (Elk Grove Village, IL, U.S.A.).
Yeast Glc-6-P dehydrogenase was the product
of Roche Molecular Biochemicals (Indianapolis, IN, U.S.A.). The bicinchoninic acid reagent (BCA) Reagent and bovine serum albumin (BSA) standard for protein assays were
purchased from Pierce Chemical Co.
(Rockford, IL, U.S.A.). AffiGel Blue was from
BioRad Laboratories (Richmond, CA, U.S.A.)
and silica gel 60 plates for thin-layer chromatography were products of Merck (Darmstadt,
Germany).
Assay of hexokinase activity and protein.
Hexokinase activity was determined using a
previously described [24] spectrophotometric
assay; Glc-6-P formation was coupled to
NADPH production, monitored at 340 nm, in
the presence of excess Glc-6-P dehydrogenase.
With samples containing Triton X-114, the assay mixtures were preincubated for approximately 2 min (to allow dissolution of the Triton X-114) before initiation of the reaction by
addition of ATP.
Protein was determined by the bicinchoninic
acid method, with BSA as standard.
Preparation of brain mitochondria. Mitochondria were prepared as described previously [25]. Briefly, brains were homogenized
in 0.25 M sucrose and a crude mitochondrial
fraction was obtained by standard centrifugal
fractionation. The crude mitochondrial pellet
was resuspended in 0.25 M sucrose, layered
on 10 ml of 1.2 M sucrose, and centrifuged for
90 min at 35 000 r.p.m. in a Beckman SW27
rotor. The mitochondrial pellets were resuspended in 0.25 M sucrose to give a protein
concentration of approx. 10 mg/ml; the specific activity of mitochondrial hexokinase was
0.4–0.8 u/mg mitochondrial protein. Mitochondrial preparations were divided into convenient aliquots and stored at –80°C.
Solubilization of mitochondrial hexokinase with Glc-6-P and with high salt. Mitochondria were diluted to a concentration of
0.2 mg mitochondrial protein/ml with 0.25 M
sucrose, 1.2 mM Glc-6-P, and 10 mM Hepes,
1047
pH 8.2. After incubation for 30 min at room
temperature, the suspensions were centrifuged at 5°C for 10 min at 20 000 ´ g. The
supernatants contained the hexokinase solubilized with Glc-6-P, i.e, hexokinase which had
been bound at Type A sites.
To obtain hexokinase bound to Type B sites,
bovine brain mitochondria were first treated
with Glc-6-P, as above. The mitochondria, now
containing only hexokinase bound to Type B
sites, were resuspended in 0.25 M sucrose, 2
mM MgCl2, 0.8% (v/v) Triton X-100, 10 mM
Hepes, pH 8.2, to give a protein concentration
of approx. 1.5 mg/ml. After 5 min on ice, the
suspension was centrifuged at 20 000 ´ g for
10 min. The resulting pellet was resuspended
in 0.25 M sucrose, 0.15 M NaCl, 20 mM
thioglycerol, 1% (v/v) Triton X-100, 10 mM
Hepes, pH 8.2, to give a protein concentration
of approx. 0.2 mg/ml. The suspension was incubated on ice for 20 min, with occasional stirring, then centrifuged at 20 000 ´ g for 10
min. The supernatant contained the saltsolubilized hexokinase, previously bound to
Type B sites.
Partial purification of hexokinase from
Type A and Type B sites of bovine brain
mitochondria. This was done using a modification of the previously described procedure
for purification of Type I hexokinase from rat
brain [24].
To the Glc-6-P-solubilized hexokinase (approx. 4 u of activity) from bovine brain mitochondria (see above), thioglycerol and EDTA
were added to final concentrations of 20 mM
and 0.5 mM, respectively, and the pH adjusted
to 7.0 with phosphoric acid. The enzyme was
applied to a small (2 ml bed volume) of AffiGel
Blue, equilibrated with 10 mM Glc, 0.5 mM
EDTA, 20 mM thioglycerol, 50 mM Tris/Cl,
pH 7.0. The column was washed extensively
with this same buffer, then with buffer of the
same composition but at pH 8.0. A final wash
with reverse flow (i.e., upward through the
column) was done with this same buffer at pH
8.0 but containing 20% (v/v) glycerol. The enzyme was eluted by continued reverse flow
1048
J. Hutny and J.E. Wilson
with the latter buffer, to which 1.5 mM Glc-6-P
was added. The eluted enzyme was concentrated in an Amicon ultrafiltration device with
YM30 membrane, diluted 25-fold with 0.25 M
sucrose, and again concentrated. This process
was repeated twice to provide the enzyme in
0.25 M sucrose, essentially free from previous
buffer components. Final yield was approximately 1 unit (25% of initial activity).
The hexokinase from Type B sites was purified by a similar procedure. The salt-solubilized enzyme (approx. 5 units), obtained as
above, was dialyzed against 0.5 mM EDTA, 10
mM Glc, 20 mM thioglycerol, 10 mM sodium
phosphate, pH 7.0. The enzyme was then purified by chromatography on AffiGel Blue as described above. Final yield was approx. 2 units
(40% of initial activity).
Phase partitioning with Triton X-114.
Mitochondrial constituents were partitioned
between detergent (hydrophobic) and aqueous phases based on the procedure of Bordier
[26]. Mitochondria were resuspended at a protein concentration of 0.5–1 mg/ml in 0.25 M
sucrose, 0.5 mM EDTA, 1% (v/v) Triton X-114,
20 mM Hepes, pH 7.4. After incubation for 20
min on ice, with occasional mixing, the samples were centrifuged at 4°C and 100 000 ´ g
for 60 min in a Beckman Ti50 rotor. Aliquots
(120 ml) of the supernatant were transferred
to prechilled 500 ml microfuge tubes on ice;
where indicated, MgCl2 was added. The samples were incubated for 2 min at 37°C, resulting in development of turbidity as the detergent and aqueous phases separated. After
centrifugation for 1 min at room temperature
in a microfuge to resolve the aqueous and hydrophobic phases, the overlying aqueous layer
was carefully removed. Following the protocol
of Bordier [26], additional buffer and detergent were added so that the volume and composition of the two fractions were similar.
Hexokinase content was then determined by
activity assay and/or quantitative densitometry after SDS gel electrophoresis. For electrophoresis, proteins were precipitated by ad-
2000
dition of three volumes of cold (–18°C) acetone and collected by centrifugation.
Electrophoresis and densitometry. SDS
gel electrophoresis was done on 6.5–20%
acrylamide gradient gels and the gels stained
with Coomassie Blue, as previously described
[27]. Gels were imaged with a GDS-2000 gel
documentation system (UVP Inc., San Gabriel, CA, U.S.A.) and band intensity determined with the SW-2000 quantitative image
analysis software provided by the manufacturer.
Extraction and analysis of mitochondrial lipids. Mitochondrial lipids were extracted by the procedure of Bligh and Dyer
[28] and analyzed by two-dimensional thinlayer chromatography on silica gel plates using solvent “System A” of Rouser & Fleischer
[29]: 1st dimension, chloroform/methanol/
water (65 : 25 : 4, by vol.); 2nd dimension,
n-butanol/acetic acid/water (60 : 20 : 20, by
vol.). Identification of lipids was based on
co-migration with authentic standards as well
as detection with reagents specific for particular lipids [30]. Spots representing the most
abundant lipids were scraped from the plates,
the lipids eluted with chloroform, and samples
analyzed by fast atom bombardment tandem
mass spectrometry in the Michigan State University Mass Spectrometry Facility. This provided confirmation of the identity, as well as
additional information about the nature of
acyl groups present.
Enrichment of bovine brain mitochondria in specific phospholipids. Bovine brain
mitochondria were enriched in particular
phospholipids with minor modifications of the
procedure of Parlo & Coleman [31]. “Nude”
Sephadex G-10 beads were prepared by washing the beads with chloroform/methanol (2 : 1,
v/v) followed by drying under vacuum. Portions (100 mg) of the nude beads were placed
in 25 ml round bottom flasks and mixed with
5 ml of a 1 mM solution of the indicated
phospholipid in chloroform/methanol (2 : 1,
v/v). The solvent was then removed on a ro-
Vol. 47
Mitochondrial binding sites for hexokinase
tary
evaporator,
leaving
the
phospholipid-coated Sephadex beads deposited as a thin film at the periphery of the flask.
Mitochondrial suspension (2 ml) containing
8–10 mg mitochondrial protein in 0.25 M sucrose, 5 mM Hepes, pH 7.4, was added. The
flasks were incubated on ice with gentle shaking for 30 min. The suspension was then layered over 30 ml of 60% (w/v) sucrose and centrifuged in a clinical centrifuge (approx. 400
´ g) for 10 min. Mitochondria were carefully
harvested from the interface above the 60%
sucrose, diluted with 0.25 M sucrose and collected by centrifugation at 20 000 ´ g for 10
min, and finally resuspended in 0.25 M sucrose. Lipid extracts were examined by
two-dimensional thin-layer chromatography,
which confirmed enrichment (estimated 2- to
3-fold, based on intensity of the spot) in the indicated phospholipid.
Statistical analysis. Data were analyzed using GraphPAD Instat Version 1.13 (Graph
Pad Software, San Diego, CA, U.S.A.).
RESULTS
Partitioning of mitochondrial hexokinase
between aqueous and hydrophobic phases
Proportionally more white matter is in the
brains of larger than in smaller animals, and
this part of brain tissue is almost depleted of
mitochondria. Thus to obtain comparable
yield of mitochondria from brains of different
origin, bigger chunks of white matter were removed from bovine brain before its storage in
liquid nitrogen (see “Material and Methods”).
This small difference in the treatment of bovine material did not change the hexokinase
distribution between the A and B sites. Approximately 40% release of hexokinase by
Glc-6-P was found by Kabir & Wilson [18] in
mitochondria prepared from the whole bovine
brains. Very close values, usually between
36% and 43% were also observed in several experiments of this work (see Table 2). Such re-
1049
sult was expected, since only mitochondrially
bound pool of hexokinase was analysed and
most of the brain hexokinase is bound to mitochondria [4–9].
Solubilization of hexokinase bound to Type
B sites requires treatments that disrupt membrane integrity [16]. This suggested an intimate interaction between the enzyme and the
outer mitochondrial membrane, more-or-less
equivalent to that seen with “integral” membrane proteins [32]. Moreover, divalent cations have been shown to be important in mediating the interaction between hexokinase and
the mitochondrial membrane [12, 17]. Based
on these considerations, we decided to examine the partitioning of mitochondrial hexokinase between hydrophobic and aqueous
phases using the method of Bordier [26], and
the effects of divalent cations on that partitioning. This method takes advantage of the
fact that Triton X-114 is miscible with water at
low temperatures, but condenses into a discrete phase after warming to moderate temperature (about 30°C). Thus, solubilization of
membranes with Triton X-114 in aqueous
buffer at low temperatures, followed by a brief
incubation at higher temperature, results in
development of discrete aqueous and detergent phases, with “integral” membrane proteins being selectively partitioned into the hydrophobic phase while hydrophilic (peripheral) membrane proteins are found in the
aqueous phase [26].
As expected for a typical water-soluble protein [26], the purified Type 1 hexokinase from
rat brain [24] was largely confined to the aqueous phase after partitioning in either the absence or presence of 2 mM MgCl2 (Table 1).
The same was true for the hexokinase obtained from bovine brain mitochondria by elution with either Glc-6-P (from Type A sites) or
high salt (from Type B sites). When intact rat
brain mitochondria, containing hexokinase
bound predominantly to Type A sites, were
subjected to the partitioning procedure, with
or without addition of MgCl2, the hexokinase
again behaved as a hydrophilic protein, with
1050
J. Hutny and J.E. Wilson
2000
Table 1. Partitioning of hexokinase between aqueous and hydrophobic (Triton X-114) phases
Percent of hexokinase activity in aqueous phase
Sample partitioned
Originally bound to
No MgCl2
+2 mM MgCl2
Purified hexokinase from
rat brain
Predominantly Type A sites
91
92
Purified Glc-6-P-solubilized
hexokinase, bovine brain
Type A sites
94
88
Purified salt-solubilized
hexokinase, bovine brain
Type B sites
86
83
Predominantly Type A sites
85 (76)
82 (77)
Type A and Type B sites
87 (80)
37 (39)
Rat brain mitochondria
b
Bovine brain mitochondria
b
a
a
Hexokinase activity was determined for both aqueous and detergent phases; total activity recovered was > 95% of initial activity.
For these samples, the distribution between aqueous and detergent phases was also determined by densitometric analysis after
SDS gel electrophoresis, yielding the values shown in parenthesis.
b
the enzyme being found predominantly in the
aqueous phase. This was also true when isolated bovine brain mitochondria, containing
hexokinase bound to both Type A and Type B
sites, were partitioned in the absence of divalent cation. But when partitioning of bovine
brain mitochondria was done in the presence
of 2 mM MgCl2, there was a marked decrease
in the amount of hexokinase found in the
aqueous phase, with corresponding increase
in the amount in the hydrophobic phase.
The proteins present in aqueous and detergent phases after partitioning of rat or bovine
brain mitochondria were examined by SDS
gel electrophoresis, and hexokinase band was
identified by comparison with purified rat
brain hexokinase used as a marker (Fig. 1). As
in previous studies [33], a slight difference in
mobility of hexokinase from bovine and rat
mitochondria has been noticed. In a separate
experiment the identity of the band designated as type I hexokinase was confirmed in
both bovine and rat SDS extracts by immunoblotting with antihexokinase I (not shown).
Electrophoresis permitted several observations of interest. First, it was apparent that
the overall protein composition of bovine and
rat brain was similar; most of the components
seen in rat brain mitochondria had their counterparts in bovine brain mitochondria, and
vice versa. This is, of course, not unexpected
since mitochondria, whatever their source,
must contain the basic set of proteins required for common mitochondrial functions.
Second, despite the fact that addition of
MgCl2 did not affect the partitioning of rat
brain mitochondrial hexokinase (Table 1), it
was clear that partitioning of several other rat
brain mitochondrial proteins into the hydrophobic phase was increased by the inclusion of
Figure 1. Electrophoretic analysis of proteins
present in detergent and aqueous phases after
Triton X-114 partitioning of rat and bovine brain
mitochondria.
Lane 1 is purified rat brain hexokinase [24], run as a
marker. Lanes 2–9 contained detergent (D) or aqueous
(Aq) phases after Triton X-114 partitioning [26] of rat
(R) or bovine (B) brain mitochondria in the absence (–)
or presence (+) of 2 mM MgCl2, as described in
Methods.
Vol. 47
Mitochondrial binding sites for hexokinase
this salt, while other protein components, together with hexokinase, remained in the aqueous phase in either the absence or presence of
MgCl2. Thus, the effect of MgCl2 was selective.
Third, a similar effect was seen with bovine
brain mitochondria, with inclusion of MgCl2
during the partitioning resulting in a selective
shift of several proteins into the hydrophobic
phase. Moreover, several of the bovine brain
mitochondrial proteins showing increased
partitioning into the detergent phase were virtually identical in mobility to rat brain mitochondrial proteins showing similar behavior;
these are reasonably presumed to represent
homologs, and would thus indicate that these
proteins show similar partitioning behavior,
whether derived from rat or bovine brain mitochondria. Thus, the difference in partitioning of hexokinase is not reflective of a general
difference in partitioning behavior of proteins
from rat and bovine brain mitochondria.
It should be apparent that partitioning of mitochondrial proteins must be dependent on
factors more complex than a simple division
into “integral” or “peripheral” proteins. In the
study of Bordier [26], partitioning of integral
membrane proteins into the detergent phase
was attributed to a “hydrophobic domain” expected to be included within the structure of
such proteins and responsible for interaction
with the hydrophobic core of the membrane.
The hydrophobic domain would, of course, be
intrinsic and hence partitioning into the detergent phase expected to be independent of
the presence or absence of MgCl2. This is in
contrast to the observed MgCl2-induced shift
of several proteins (including the hexokinase
of bovine, but not rat, brain mitochondria) to
the detergent phase.
The fact that an increased proportion of several proteins were found to partition into the
detergent phase in the presence of MgCl2 suggested the possibility that inclusion of MgCl2
had induced the formation of an aggregate. Indeed, extended ultracentrifugation (100 000
´ g for 30 min at room temperature in a
Beckman Airfuge) of the detergent phase gave
1051
a small pellet. Analysis of the pellet and the
overlying detergent phase by SDS gel electrophoresis showed them to be indistinguishable
in protein composition (and as shown in
Fig. 1), with approximately equal amounts of
protein being found in both pellet and “soluble” in the detergent phase. Thus, it would appear that at least some aggregation may be associated with the observed partitioning behavior.
Mg++-induced partitioning into the detergent phase is inversely related to the proportion of Type A sites present in the mitochondria
The proportion of hexokinase bound to Type
A sites of brain mitochondria increases in the
order of bovine < rabbit < guinea pig < rat
brain (Table 2). This is in agreement with previous findings [16] except that the modest difference between rabbit and guinea pig brain
mitochondria had not been noted in the earlier work. When partitioning of hexokinase
bound to these mitochondria was examined, it
was found that the Mg++-induced partitioning
into the detergent phase decreased in a similar
order (Table 2), with values for bovine and rat
brain mitochondria lying at the extremes
while values for guinea pig and rabbit (not significantly different from each other) were intermediate.
Mg++-induced partitioning into the detergent phase is not associated with occupancy
of Type B sites
The results in Table 2 suggested the possibility that it was hexokinase bound at Type B
sites that was selectively partitioned into the
detergent phase in the presence of Mg++.
Thus, as the proportion of hexokinase bound
to Type A sites increased, and the proportion
occupying Type B sites correspondingly decreased, decreased partitioning into the detergent phase was observed. If this were the case,
then elution of the enzyme bound to Type A
1052
J. Hutny and J.E. Wilson
sites prior to partitioning (i.e., all of the
hexokinase left on the mitochondria would be
occupying Type B sites) should result in a
marked increase in the hexokinase found in
the detergent phase. However, when partitioning was done with bovine brain mitochondria which had been depleted of hexokinase
bound at Type A sites by incubation with
Glc-6-P, the Mg++-induced increase in partitioning into the detergent phase was 27 ± 3%
2000
parting sufficient hydrophobic character to
lead to significant partitioning into the detergent phase. Previous work [34] implicating
membrane phospholipids in the hexokinasemitochondrial interaction prompted experiments to examine possible involvement of
phospholipids in the observed partitioning behavior.
In an initial experiment, rat and bovine
brain mitochondria were solubilized with Tri-
Table 2. Type A sites and partitioning of brain mitochondria from different species
Hexokinase activity
a
++
% solubilized with Glc-6-P
b
(Type A sites)
Mg -induced decrease in aqueous
c
phase
Bovine
38 ± 2 (5)
30 ± 3 (8)
Rabbit
60 ± 3 (3)
18 ± 4 (7)
Guinea pig
72 ± 3 (4)
23 ± 4 (7)
Rat
85 ± 2 (5)
6 ± 3 (8)
Brain mitochondria
a
b
Mean ± S.D. for number of independent experiments shown in parenthesis. Values for rabbit and guinea pig are significantly
c
different with P < 0.01. All other comparisons are significantly different with P < 0.001. Values shown are percent of hexokinase
found in the aqueous phase in the absence of MgCl2 minus the percent found in the aqueous phase after partitioning in the presence of MgCl2. Values for rabbit and guinea pig are not significantly different (P > 0.05). Values for bovine and guinea pig are significantly different with P < 0.05. All other comparisons are significantly different with P < 0.001.
(for four independent experiments), not significantly different from the value of 30 ± 3%
(Table 2) obtained with bovine brain mitochondria as isolated (i.e., with hexokinase occupying both Type A and Type B sites). Thus,
the partitioning results depend on the relative
proportions of Type A and Type B sites present on the mitochondria, but are not determined by whether the hexokinase being partitioned is actually occupying Type A or Type B
sites.
Partitioning of hexokinase into the detergent phase is dependent on phospholipids
Since hexokinase itself did not partition into
the detergent phase (Table 1), these results
suggested that partitioning behavior was dependent on Mg++-induced interaction of
hexokinase with one or more hydrophobic
components of bovine brain mitochondria, im-
ton X-114, as for a standard partitioning experiment (see Methods). However, prior to the
brief incubation at 37°C to produce the phase
separation, 50 milliunits of phospholipase C
were added and the samples were incubated
for 3 h at 5–10°C; control samples were incubated in the same way, but without addition of
phospholipase C. The samples were then
warmed to 37°C and partitioning continued as
usual. The slight Mg++-induced increase in
partitioning of rat brain mitochondrial hexokinase was virtually unchanged by phospholipase C treatment, being 7% and 4% in the
control and treated samples, respectively;
these values are in good agreement with those
seen in standard partitioning experiments
(Table 2). In contrast, the corresponding values for bovine brain mitochondria were 26%
(again in agreement with results in standard
partitioning protocol — Table 2) and 5% for
control and treated samples, respectively.
Vol. 47
Mitochondrial binding sites for hexokinase
The effect of phospholipase was dependent
on both time of treatment and amount of
phospholipase used. Thus, after solubilization
of bovine brain mitochondria with Triton
X-114, as above, the Mg++-induced increase in
partitioning into the detergent phase was 24%
in the control (no phospholipase added) and
15% and 7% after 1 and 2 h, respectively, treatment with 100 milliunits of phospholipase C.
In a similar experiment, the detergentsolubilized bovine brain mitochondria were
incubated for 1 h with 0, 50, 100, or 200
milliunits of phospholipase C; subsequent partitioning showed the Mg++-induced increase in
partitioning into the detergent phase was 25,
14, 7 and 4%, respectively.
Phospholipase treatment of intact mitochondria produced similar results. Thus, bovine
brain mitochondria were incubated (5–10°C,
30 min) with 0, 40, or 200 milliunits of phospholipase C, then collected by centrifugation.
Assay of the supernatant revealed < 5% of the
hexokinase had been released during phospholipase treatment. The mitochondrial pellets were resuspended and a portion used for
partitioning by the usual protocol. The
Mg++-induced increase in partitioning into the
1053
detergent phase was 31, 26, and 14% for the
mitochondria treated with 0, 40, and 200
milliunits of phospholipase C, respectively.
Lipids were extracted from another portion of
the phospholipase-treated mitochondria; analysis by 2-dimensional thin-layer chromatography confirmed substantial decrease in all
phospholipid components.
In all of the above experiments, proteins
present in aqueous and detergent phases were
examined by SDS gel electrophoresis. Control
samples (no phospholipase added) gave results as shown in Fig. 1, lanes 6–9. Phospholipase treated samples were, qualitatively,
identical to controls, but it was evident that
phospholipase treatment resulted in a decreased partitioning of all protein components into the detergent phase, i.e., the effects
of phospholipase were not selective for hexokinase.
Comparison of the phospholipid composition of bovine and rat brain mitochondria
Lipids extracted from bovine and rat brain
mitochondria were examined by 2-dimensional thin-layer chromatography (Fig. 2). The
Figure 2. Thin-layer chromatographic analysis of lipids from rat and bovine brain mitochondria.
Lipids were extracted and analyzed by thin-layer chromatography as described in Methods. Prominent components
were phosphatidylethanolamine (PE), two spots (PC1 and PC2) identified as phosphatidylcholine with differing acyl
group composition (see text), and two spots (PI/PS1 and PI/PS2), both of which contained phosphatidylinositol as
well as phosphatidylserine. Sphingomyelin (Sph) was detected in extracts from bovine brain mitochondria, but not
rat brain mitochondria. Both types of mitochondria also contained sulfatide (unmarked spot, upper left from
“PI/PS1” label).
1054
J. Hutny and J.E. Wilson
same major lipid components were detected
in both kinds of mitochondria but there were
some differences, reproducibly seen in several
such comparisons.
First, phosphatidylcholine was partially resolved into two overlapping spots, designated
PC1 and PC2 in Fig. 2. PC1 was consistently
more intense in extracts from rat brain mitochondria, while the more diffuse, slower moving PC2 was more intense with bovine brain
mitochondria. Mass spectrometric analysis
confirmed identification as phosphatidylcholine, but revealed that the PC1 and PC2
differed in fatty acid composition. PC1 contained primarily stearoyl (18 : 0) and 20 carbon acyl groups, the latter with 1, 2, or 4 double bonds. In contrast, PC2 contained shorter
acyl groups, primarily palmitoyl (16 : 0), with
some stearoyl (18 : 0), oleoyl (18 : 1), and
myristoyl (14 : 0) groups.
Secondly, sphingomyelin (Sph in Fig. 2) was
consistently detected as a minor component
in extracts from bovine brain mitochondria,
but was barely detectable in rat brain mitochondrial extracts.
Finally, two spots (labelled PI/PS1 and PI/
PS2 in Fig. 2) migrated in the region expected
for phosphatidylinositol and phosphatidylserine. The relative amounts of PI/PS1 and
2000
PI/PS2 were comparable in rat brain mitochondria, while PI/PS2 < PI/PS1 in bovine
brain mitochondria. Although well resolved,
both spots were ninhydrin-positive, indicating
the presence of phosphatidylserine but this
could not be confirmed by mass spectrometry.
Mass spectrometry did confirm the presence
of phosphatidylinositol, with stearoyl (18 : 0)
and arachidonoyl (20 : 4) as primary acyl
groups, in both spots. Commercially obtained
phosphatidylserine and phosphatidylinositol
both gave a single large spot that encompassed both the PI/PS1 and PI/PS2 regions.
The basis for resolution of the mitochondrial
phospholipids into two spots, with both
phosphatidylinositol and phosphatidylserine
being present in both spots, is unclear.
Solubilization of mitochondrial hexokinase
with Triton X-100, Triton X-114, and
digitonin
The mitochondrial hexokinases from bovine,
rabbit, guinea pig, and rat differed in their
susceptibility to solubilization by Triton X-100
and Triton X-114 (Table 3), with the bovine enzyme being most resistant, the rat brain enzyme most susceptible, and the rabbit and
guinea pig enzymes intermediate. In each
Table 3. Solubilization of mitochondrial hexokinase with Triton X-114, Triton X-100, or digitonin
% of hexokinase solubilized with brain mitochondria from
(mg/mg
Detergent added Concentration
mitochondrial protein) Bovine
Rabbit
Guinea pig
Rat
None
–
Triton X-114
Triton X-100
Digitonin
a
4±2
8±2
8±3
9±2
0.2
10 ± 2
19 ± 3
18 ± 2
35 ± 5
0.35
29 ± 1
48 ± 7
47 ± 2
71 ± 5
0.5
68 ± 3
87 ± 4
85 ± 1
92 ± 4
1.0
88 ± 4
91 ± 2
95 ± 1
96 ± 1
0.4
30 ± 4
47 ± 7
43 ± 10
59 ± 8
2
50 ± 4
70 ± 4
69 ± 11
78 ± 6
50
89 ± 1
93 ± 2
95 ± 3
96 ± 3
0.1
6±2
13 ± 1
12 ± 4
13 ± 3
0.8
7±2
14 ± 3
12 ± 1
17 ± 6
1.6
13 ± 2
19 ± 3
18 ± 5
30 ± 4
a
Values shown are mean ± S.D. for nine samples with no added detergent, and for three samples at each concentration of detergent.
Vol. 47
Mitochondrial binding sites for hexokinase
case, Triton X-114 was more effective than
Triton X-100. These detergents differ in their
relative hydrophobic character, with Triton
X-100 containing an average of 9.5 hydrophilic oxyethylene groups per (hydrophobic)
octylphenol headgroup, while the more hydrophobic Triton X-114 contains only 7–8 oxyethylene groups per octylphenol (Product Information Sheet, Sigma Chemical Co.).
Confirming previous results [10, 35, 36],
brain mitochondrial hexokinase was rather resistant to solubilization by digitonin (Table 3).
Again, there were differences seen with brain
mitochondria from different species, with
hexokinase of bovine brain mitochondria being the most resistant, the hexokinase of rat
brain mitochondria most susceptible, and
hexokinase of guinea pig and rabbit mitochondria intermediate in susceptibility to solubilization with digitonin. As in previous studies [10, 35, 36], a maximum of about 30% of
the hexokinase could be solubilized from rat
brain mitochondria; with bovine brain mitochondria, a maximum of about 15% could be
solubilized at the highest digitonin concentrations used. In contrast, 80–90% of the adenylate kinase, a marker for the intermembranal space, was released from both bovine
and rat brain mitochondria with only about
0.5 mg digitonin/mg protein ([10, 35, 36], and
data not shown). Digitonin is thought to exert
1055
its membrane-disrupting effect through complexing with cholesterol [37]. Thus these results are consistent with the suggestion [35]
that the majority of the hexokinase on brain
mitochondria is associated with cholesteroldeficient regions of the outer membrane
thought to be located at contact sites [35,
38–40].
Effect of enrichment of mitochondrial membranes in specific phospholipids
Bovine brain mitochondria were selectively
enriched in specific phospholipids and then
treated with Glc-6-P to release hexokinase
from Type A sites. Subsequent re-incubation
of the mitochondria with Glc-6-P evoked no
further Glc-6-P-dependent release — as expected, since the enzyme had previously been
removed from Type A sites. Enrichment of the
mitochondria in the indicated phospholipids
had, at most, a marginal effect on solubilization of hexokinase in either the absence or
presence of Glc-6-P (Table 4). However, if the
enriched mitochondria were first incubated at
37°C, solubilization in both the absence and
presence of Glc-6-P was enhanced by enrichment in acidic phospholipids while enrichment in phosphatidylcholine or phosphatidylethanolamine was without significant effect. These results are consistent with previ-
Table 4. Enrichment in acidic phospholipids, but not phosphatidylcholine or phosphatidylethanola
amine, alters binding of hexokinase to bovine brain mitochondria
No incubation at 37°C
% hexokinase solubilized
Mitochondria enriched in
No Glc-6-P
b
Plus Glc-6-P
After 4 h incubation at 37°C
% hexokinase solubilized
No Glc-6-P
Plus Glc-6-P
No exogenous lipids
7±2
9±3
9±0
19 ± 5
Phosphatidylcholine
6±2
8±2
10 ± 2
16 ± 4
Phosphatidylethanolamine
6±1
8±1
9±3
15 ± 4
Phosphatidylserine
11 ± 1
15 ± 2
28 ± 8
40 ± 6
Phosphatidylinositol
13 ± 8
14 ± 5
28 ± 4
42 ± 12
a
Bovine brain mitochondria were enriched in the indicated phospholipids; as a control (“no exogenous lipids”), mitochondria
were subjected to the same enrichment protocol but without addition of exogenous lipids. The mitochondria were treated with
Glc-6-P (as in Methods) to elute hexokinase bound at Type A sites, then either maintained on ice or incubated at 37°C for 4 h.
Aliquots were then incubated for 20 min at room temperature either with or without addition of 1.2 mM Glc-6-P. Following
b
centrifugation, the percent of total hexokinase activity in solubilized form (i.e., in the supernatant) was determined. Values
shown are mean ± S.D.; n = 5 for control samples (no exogenous lipids); n = 3 for all others.
1056
J. Hutny and J.E. Wilson
ous work [34] implicating acidic phospholipids as significant factors influencing the
hexokinase-mitochondrial membrane interaction, and suggest that reorganization of
phospholipid components (during the incubation at 37°C) was required for manifestation
of this effect.
Another interesting aspect of the results
shown in Table 4 should be noted. As mentioned above, if control (i.e., not enriched in
exogenous lipids) were pre-treated with
Glc-6-P to elute hexokinase from the Type A
sites and then kept at 0°C before a second incubation with Glc-6-P, no further Glc-6-P-dependent release was seen. However, if the mitochondria were incubated at 37°C prior to
the second treatment with Glc-6-P, there was a
highly significant (P < 0.003) increase in
Glc-6-P-dependent solubilization but no effect
on solubilization in the absence of Glc-6-P. In
other words, some redistribution of the enzyme, initially bound at Type B sites, had occurred during the incubation at 37°C.
Whether this resulted from spontaneous dissociation of the enzyme from Type B sites and
rebinding at Type A sites, or from an actual
conversion of some of the Type B sites to Type
A sites cannot be determined from these results. In either case, these results suggest a
dynamic relationship between Type A and
Type B sites.
DISCUSSION
The present study has extended previous
work [16, 18] indicating notable differences in
the binding of hexokinase to brain mitochondria from different species. It appears that the
lipid environment may play a significant role
in determining the properties of the binding
site for hexokinase in mitochondrial membrane. Authors’ intention was to use the outer
mitochondrial membrane as a basic material
for experiments of this work. But the separation methods of mitochondrial membranes
from brain tissue are still not developed
2000
enough to obtain the purified product. For unknown reason the “swelling and shrinking”
procedure, which give excellent results with
other tissues, does not work well with brain
mitochondria. Also alternative procedures
[35, 41] do not provide a satisfactory enrichment of standard outer membrane markers,
and our attempts of their improvement were
unsuccessful. Because of these technical problems, we limited the analysis to whole mitochondria.
A general resistance of mitochondrially
bound hexokinase to solubilization by digitonin was a common characteristic for all species examined, yet there were differences implying that the relative cholesterol content —
and hence susceptibility to disruption by digitonin [37] — varied with species. It is clear that
this does not represent a general resistance of
the outer mitochondrial membrane to digitonin since the membrane is disrupted, as indicated by release of adenylate kinase from
the intermembranal space as well as solubilization of other outer mitochondrial membrane enzymes [10, 35, 36] at digitonin concentrations that cause release of only marginal amounts of hexokinase. That “lipid domains”, i.e., regions of varying lipid composition, may exist within a membrane was first
proposed by Karnovsky et al. [42] and is now a
generally accepted principle of membrane
structure [43, 44]. Indeed, formation of such
domains may be induced by membrane proteins [45]. The heterogeneous response of various outer mitochondrial membrane enzymes
to solubilization by digitonin implies that they
reside in domains differing in relative content
of cholesterol, and likely other lipids. The apparently cholesterol-deficient regions, in
which most of the mitochondrially bound
hexokinase is located, are suggested to be contact sites [35, 38–40], i.e., regions of intimate
contact, perhaps fusion, between the inner
and outer mitochondrial membranes.
Approximately 70% of the hexokinase bound
to rat brain mitochondria ([10, 35, 36], and
present work) was resistant to solubilization
Vol. 47
Mitochondrial binding sites for hexokinase
by digitonin, while this was about 90% for the
bovine mitochondrial hexokinase, with mitochondrial hexokinases from guinea pig and
rabbit brain falling between these extremes.
While the Triton detergents were much more
effective than digitonin, an analogous order
was seen in the relative susceptibility to
solubilization; in all cases Triton X-114 was
more effective (on a mg detergent per mg protein basis) that the more hydrophilic Triton
X-100. These results thus imply that solubilization of hexokinase requires access to a
highly hydrophobic cholesterol-poor domain,
with relative accessibility and cholesterol content decreasing in the order rat > guinea pig ?
rabbit > bovine.
It is clear that phospholipid composition,
particularly content of acidic phospholipids,
can markedly affect the interaction of hexokinase with mitochondria ([34], and present
work). However, as shown by the results in Table 4, incorporation of acidic phospholipids
into the mitochondrial membranes is not, in
itself, sufficient to alter the hexokinase-membrane interaction. Rather, the effects of altered phospholipid content became evident
only after preincubation of the enriched mitochondria at elevated temperature (37°C). This
suggests that the effects on hexokinase binding are seen only after a reorganization of the
phospholipids, randomly incorporated during
the enrichment process, into domains closely
associated with the hexokinase binding sites.
We hypothesize that Type A and Type B sites
reside within lipid domains of differing phospholipid composition. In particular, we suggest that the lipid environment of the Type A
site may be enriched in acidic phospholipids
which have previously been shown [34] to enhance susceptibility of the mitochondrially
bound enzyme to solubilization with Glc-6-P.
No gross differences in phospholipid composition, including the acidic phospholipids, were
detected in mitochondria from rat and bovine
brain, despite the fact that these mitochondria differ considerably in relative proportion
of Glc-6-P-sensitive (Type A) sites present.
1057
However, this does not preclude the possibility that the organization of these phospholipids may differ, with increased sequestration of acidic phospholipids into “Type A”
lipid domains leading to increased proportion
of the hexokinase bound in a Glc-6-P-sensitive
manner. What would prompt such different
lipid organization in membranes that did not
differ markedly in overall lipid composition is,
of course, a major question, for which we presently cannot suggest an answer.
As noted above, the Mg++-dependent partitioning into the Triton X-114 phase is not correlated with occupancy of Type B sites but
rather with the relative proportion of Type B
sites present on the mitochondria subjected to
the partitioning regimen. Since the partitioning behavior does not depend on whether the
hexokinase is associated with both Type A and
Type B sites (bovine brain mitochondria, as
isolated) or only with Type B sites (after treatment with Glc-6-P) prior to detergent treatment, we suggest that the Mg++-dependent interaction of hexokinase with a hydrophobic
mitochondrial component (or components)
occurs after dissolution of the mitochondria
with Triton X-114; interactions within the homogeneous milieu would not be dependent on
the status of hexokinase prior to dissolution.
The sensitivity of partitioning behavior to
phospholipase treatment implies that one or
more phospholipids are involved in conferring hydrophobic character on hexokinase,
and the correlation between proportion of
Type B sites and partitioning behavior further
implies that the relative amount of the responsible phospholipid(s) is similarly correlated
with Type B sites. Gross differences in phospholipid composition of bovine and rat brain
mitochondria were not seen, but it is certainly
conceivable that interaction with hexokinase
may depend on more subtle factors. Unfortunately, attempts to identify specific phospholipids that partition with the hexokinase
into the detergent phase are doomed since all
hydrophobic components, associated with
hexokinase or otherwise, are found in the hy-
1058
J. Hutny and J.E. Wilson
drophobic phase. Thus, further studies on this
question will require an alternative experimental approach.
REFERENCES
1. Katzen, H.M. (1967) The multiple forms of
mammalian hexokinase and their significance
to the action of insulin. Adv. Enzyme Regul. 5,
335–356.
2. Wilson, J.E. & Felgner, P.L. (1977) An inverse
relation between mitochondrial hexokinase
content and phosphoglucomutase activity of
rat tissues. Mol. Cell. Biochem. 18, 39–47.
3. Clarke, D.D. & Sokoloff, L. (1994) Circulation
and energy metabolism in the brain; in Basic
Neurochemistry (Siegel, G.J., Agranoff, B.W.,
Albers, R.W. & Molinoff, P.B., eds.) pp. 645–
680, Raven Press, New York.
4. Crane, R.K. & Sols, A. (1953) The associacion
of hexokinase with particulate fractions of
brain and other tissue homogenates. J. Biol.
Chem. 203, 273–292.
5. Johnson, M.K. (1960) The intracellular distribution of glycolytic and other enzymes in
rat-brain homogenates and mitochondrial preparations. Biochem. J. 77, 610–618.
6. Schwartz, G.P. & Basford, R.E. (1967) The isolation and purification of solubilized hexokinase from bovine brain. Biochemistry 6,
1070–1079.
7. Bachelard, H.S. (1967) The subcellular distribution and properties of hexokinase in the
guinea-pig cerebral cortex. Biochem. J. 104,
286–292.
8. Redkar, V.D. & Kenkare, U.W. (1972) Bovine
brain mitochondrial hexokinase. Solubilisation, purification, and role of sulfhydryl residues. J. Biol. Chem. 247, 7576–7584.
9. Magnani, M., Serafini, G., Stocchi, V., Bossù,
M. & Dacha, M. (1982) Solubilisation, purification, and properties of rabbit brain hexo-
2000
kinase. Arch. Biochem. Biophys. 216, 449–
454.
10. BeltrandelRio, H. & Wilson, J.E. (1992) Coordinated regulation of cerebral glycolytic and
oxidative metabolism, mediated by mitochondrially bound hexokinase dependent on intramitochondrially generated ATP. Arch. Biochem. Biophys. 296, 667–677.
11. de Cerqueira Cesar, M. & Wilson, J.E. (1998)
Further studies on the coupling of mitochondrially bound hexokinase to intramitochondrially compartmented ATP, generated
by oxidative phosphorylation. Arch. Biochem.
Biophys. 350, 109–117.
12.Felgner, P.L., Messer, J.L. & Wilson, J.E.
(1979) Purification of a hexokinase-binding
protein from the outer mitochondrial membrane. J. Biol. Chem. 254, 4946–4949.
13. Fiek, C., Benz, R., Roos, N. & Brdiczka, D.
(1982) Evidence for identity between the
hexokinase-binding protein and the mitochondrial porin in the outer membrane of rat liver
mitochondria. Biochim. Biophys. Acta 688,
429–440.
14. Lindén, M. Gellerfors, P. & Nelson, B.D.
(1982) Pore protein and the hexokinase-binding protein from the outer membrane of rat
liver mitochondria are identical. FEBS Lett.
141, 189–192.
15. Colombini, M., Blachly-Dyson, E. & Forte, M.
(1996) VDAC, a channel in the outer mitochondrial membrane; in Ion Channels (Narahishi,
T., ed.) vol. 4, pp. 169–202, Plenum Press,
New York.
16. Kabir, F. & Wilson, J.E. (1993) Mitochondrial
hexokinase in brain of various species: Differences in sensitivity to solubilisation by glucose-6-phosphate. Arch. Biochem. Biophys.
300, 641–650.
17. Felgner, P.L. & Wilson, J.E. (1977) Effect of
neutral salts on the interaction of rat brain
hexokinase with the outer mitochondrial
Vol. 47
Mitochondrial binding sites for hexokinase
membrane. Arch. Biochem. Biophys. 182,
282–294.
18. Kabir, F. & Wilson, J.E. (1994) Mitochondrial
hexokinase in brain: Coexistence of forms differing in solubilisation by glucose-6-phosphate
on the same mitochondria. Arch. Biochem.
Biophys. 310, 410–416.
19. Tolani, A.J. & Talwar, G.P. (1963) Differential
metabolism of various brain regions: Biochemical heterogeneity of mitochondria. Biochem.
J. 88, 357–362.
20.Neidle, A., Van den Berg, C.J. & Grynbaum, A.
(1969) The heterogeneity of rat brain mitochondria isolated on continuous sucrose gradients. J. Neurochem. 16, 225–234.
21. Blokhuis, G.G.D. & Veldstra, H. (1970) Heterogeneity of mitochondria in rat brain. FEBS
Lett. 11, 197–199.
22.Wilson, J.E. (1972) The localization of latent
brain hexokinase on synaptosomal mitochondria. Arch. Biochem. Biophys. 150, 96–104.
23.De Pinto, V., Ludwig, O., Krause, J., Benz, R.
& Palmieri, F. (1987) Porin pores of mitochondrial outer membranes from high and low
eukaryotic cells: Biochemical and biophysical
characterization. Biochim. Biophys. Acta 894,
109–119.
24. Wilson, J.E. (1989) Rapid purification of mitochondrial hexokinase from rat brain by a single affinity chromatography step on Affi-Gel
Blue. Prep. Biochem. 19, 13–21.
25.Tuttle, J.P. & Wilson, J.E. (1970) Rat brain
hexokinase: A kinetic comparison of soluble
and particulate forms. Biochim. Biophys. Acta
212, 185–188.
26.Bordier, C. (1981) Phase separation of integral membrane proteins in Triton X-114 solution. J. Biol. Chem. 256, 1604–1607.
27. Needels, D.L. & Wilson, J.E. (1983) The identity of hexokinase activities from mitochondrial and cytoplasmic fractions of rat brain
homogenates. J. Neurochem. 40, 1134–1143.
1059
28.Bligh, E.G. & Dyer, W.J. (1959) A rapid
method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37, 911–917.
29.Rouser, G. & Fleischer, S. (1967) Isolation
characterisation, and determination of polar
lipids of mitochondria. Methods Enzymol. 10,
385–406.
30.Skipski, V.P. & Barclay, M. (1969) Thin-layer
chromatography of lipids. Methods Enzymol.
14, 530–598.
31. Parlo, R.A. & Coleman, P.S. (1984) Enhanced
rate of citrate export from cholesterol-rich
hepatoma mitochondria. The truncated Krebs
cycle and other metabolic ramifications of mitochondrial membrane cholesterol. J. Biol.
Chem. 259, 9997–10003.
32.Singer, S.J. & Nicolson, G.L. (1972) The fluid
mosaic model of the structure of cell membranes. Science 175, 720–731.
33.Ureta, T., Smith, A.D. & Wilson, J.E. (1986)
Hexokinase A from mammalian brain: Comparative peptide mapping and immunological
studies with monoclonal antibodies. Arch.
Biochem. Biophys. 246, 419–427.
34. Moller, F. & Wilson, J.E. (1983) The influence
of specific phospholipids on the interaction of
hexokinase with the outer mitochondrial
membrane. J. Neurochem. 41, 1109–1118.
35.Dorbani, L., Jancsik, V., Linden, M., Leterrier,
J.F., Nelson, B.D. & Rendon, A. (1987) Subfractionation of the outer membrane of rat
brain mitochondria: Evidence for the existence of a domain containing the porin-hexokinase complex. Arch. Biochem. Biophys. 252,
188–196.
36.Parry, D.M. & Pedersen, P.L. (1990) Glucose
catabolism in brain. Intracellular localization
of hexokinase. J. Biol. Chem. 265, 1059–1066.
37. Nishikawa, M., Nojima, S., Akiyama, T., Sankawa, U. & Inoue, K. (1984) Interaction of
digitonin and its analogs with membrane cholesterol. J. Biochem. (Tokyo) 96, 1231–1239.
1060
J. Hutny and J.E. Wilson
38.Brdiczka, D. (1990) Interaction of mitochondrial porin with cytosolic proteins. Experientia
46, 161–167.
39.Kottke, M., Adam, V., Riesinger, I., Bremm,
G., Bosch, W., Brdiczka, D., Sandri, G. &
Panfili, E. (1988) Mitochondrial boundary
membrane contact sites in brain: Points of
hexokinase and creatine kinase location, and
control of Ca2+ transport. Biochim. Biophys.
Acta 935, 87–102.
40.Brdiczka, D. (1991) Contact sites between mitochondrial envelope membranes. Structure
and function in energy- and protein-transfer.
Biochim. Biophys. Acta 1071, 291–312.
41. Craven, P.A., Goldblatt, P.J. & Basford, R.E.
(1969) Brain hexokinase. The preparation of
inner and outer mitochondrial membranes.
Biochemistry 8, 3525–3532.
2000
42.Karnovsky, M.J., Kleinfeld, A.M., Hoover,
R.L. & Klausner, R.D. (1982) The concept of
lipid domains in membranes. J. Cell Biol. 94,
1–6.
43.Hinderliter, A.K., Huang, J. & Feigenson,
G.W. (1994) Detection of phase separation in
fluid phosphatidylserine/phosphatidylcholine
mixtures. Biophys. J. 67, 1906–1911.
44.Brown, D.A. & London, E. (1997) Structure of
detergent-resistent membrane domains: Does
phase separation occur in biological membranes? Biochem. Biophys. Res. Commun. 240,
1–7.
45.Boggs, J.M., Wood, D.D., Moscarello, M.A. &
Papahadjopoulos, D. (1977) Phase separation
of acidic and neutral phospholipids induced by
human myelin basic protein. Biochemistry 16,
2325–2329.